There is an unmet need to develop novel biomarkers for Alzheimer?s disease (AD) that are minimally invasive and that more broadly serve as accurate indicators of the underlying pathogenic mechanisms in brain, including impaired neuronal and synaptic function, neuroinflammation, and neurodegeneration. Previous attempts to profile blood proteomes have been hindered by the relatively small number of proteins measured, difficulties linking their expression to AD brain and various disease processes, and poor replicability. The goal of this proposal is to develop a novel plasma protein biomarker platform for AD by deeply profiling the proteomes of matched brain and plasma samples from the same cases, linking the protein networks to the extensive clinical, pathological, and molecular data available in the Accelerating Medicine Partnership for AD (AMP-AD) to nominate candidate plasma protein biomarkers and validate their performance in several independent cohorts. The research will extend and leverage our contributions to the AMP-AD consortium using discovery proteomics and systems biology to generate protein co-expression networks for human post- mortem AD brain. Our initial studies have revealed biologically meaningful groups of co-expressed proteins (i.e., modules) in cortex of AD cases that strongly correlate with key traits such as diagnosis, cognition and neuropathology. The modules identify proteins and pathways involved in AD pathophysiological processes (e.g., synaptic and cytoskeletal dysfunction, inflammation, apoptosis, and others), with stronger trait- associations than in RNA expression networks, high reproducibility across all AMP-AD cohorts, and progressive change from preclinical stages to advanced AD. In the proposed research, our collaborative teams of experts from Emory, Rush, and Oxford will first extend the discovery proteomic analyses in AMP-AD using next generation mass spectrometry and aptamer arrays, increasing the depth of protein coverage several-fold, measuring ~6000 proteins in brain (Aim 1) and plasma (Aim 2) from the same control, AD, and non-AD dementia cases in the Religious Orders Study and Memory Aging Project. An integrative analysis and machine learning will be used to nominate ~100 plasma protein candidate biomarkers most strongly associated with diagnosis, and key clinical, molecular and pathological endophenotypes. The performance of the candidate biomarkers will then be assessed using >2000 samples in three independent cohorts from the MOVE-AD consortium, the European Medical Information Framework, and Dementia Platform United Kingdom (Aim 3), The results will amplify the impact of the AMP-AD and MOVE-AD consortia with rapid and full data sharing, and establish an innovative pipeline for discovery and validation of plasma proteomics biomarkers that serve as robust and reproducible indicators of AD, including the dysregulated processes that occur in brain.